SVD(Singular Value Decomposition) and PCA (Principal Component Analysis)

1. In some cases, PCA may *not* be helpful in reducing data features. Describe a situation where this might be the case. Be brief in your answer.

2. Given an image matrix A, its singular values are given by 16.25, 1.2, 0.04 and the rest all below 0.01. How many components may be needed to get a reasonable approximation to the image. You need to explain your answer.